Towards prescriptive analytics of learning strategies

Self-regulated learning (SRL) is vital for success in online learning, requiring learners to manage their learning independently. This thesis explores using learner trace data and reinforcement learning (RL) to enhance SRL support, addressing three research gaps. First, current adaptive feedback tools lack integration of SRL principles. We propose RL-based tools to analyze SRL behavior and offer personalized feedback. Second, existing SRL analysis methods struggle with real-time feedback. We show RL's potential to model SRL behavior and provide actionable guidance. Third, we highlight the impact of trace parser methodologies on SRL analysis, advocating for transparent, reliable measurement. Our research demonstrates RL's effectiveness in supporting SRL, emphasizing transparent trace parsing and proposing an RL-based approach for personalized feedback.